Speaker diarisation
Speaker diarisation: Segmentation of a recording assigning audio portions to different speakers without necessarily identifying them by name. It answers “who spoke when” and complements transcription. Overlap, noise and number of voices affect errors. This perspective connects the definition with practical behaviour in artificial intelligence systems.
In italiano: Speaker DiarizationLegge I · Embedding, feature e spazi vettorialiIt answers “who spoke when” and complements transcription. Overlap, noise and number of voices affect errors.